高容错(2,1,m)卷积码快速盲识别方法

摘要

A fast method for blind identification of (2,1,m) convolutional codes in the case of high error bit ratios is proposed to solve the problems that the existing methods for blind identification of convolutional codes have low robustness and need to know encoder parameters a priori and large computation.First,a set of parity check equations is established and its solution is the parity check vector of a (2,1,m) convolutional code.Then,based on the structural properties of coefficients of the parity check equations,unknown elements of the parity check vector are recursively estimated from values of the elements having been estimated.When there are errors in the received bits,the estimation of an unknown element is jointly determined by several parity check equations in each recursion.Finally,the correctness of estimations of the parity check vector is tested based on the property of free distance of convolutional codes and constant false-alarm criterion,and the generator matrix of the (2,1,m) convolutional code is identified accordingly.When the encoder parameters are unknown,these parameters can be quickly identified from the testing results of the parity check vector estimations.Simulation results show that the proposed method has good tolerance to error bits and low computational complexity,and does not need to know the encoder parameters a priori.When the error bit ratio is 0.08,the correct identification ratio of the method reaches 80％,while the method based on matrix analysis cannot achieve a correct identification.%针对现有的卷积码识别方法存在容错性不高,需已知编码参数且计算量大的问题,提出一种高容错的(2,1,m)卷积码快速盲识别方法.首先建立以(2,1,m)卷积码校验序列为解向量的含错校验方程组;然后基于校验方程系数的结构特性,循环利用校验序列的已知元素值递推估计其未知元素,在有误码条件下,进一步利用多个校验方程联合判决未知元素估计值,实现误码条件下校验序列的快速估计;最后基于卷积码自由距离特性及恒虚警准则检验校验序列估计值的正确性,并相应地识别出(2,1,m)卷积码生成多项式矩阵.在编码参数未知的情况下,根据校验序列估计值的检测结果,快速识别编码参数.仿真实验表明:该方法具有较高的容错性和较低的计算复杂度,无需先验已知编码参数;当误码率为0.08时,识别正确率能达到80％以上,而此时矩阵分析识别法已无法正确识别卷积码.